Smoking is not better for you than sleep apnea.

摘要

Drs. Yegneswaran and Shapiro have recently suggested here that CPAP non-compliance in obstructive sleep apnea (OSA) patients has a much greater effect on mortality than smoking does. But they have incorrectly used summary data from our Busselton Sleep Apnea cohort to support this claim. The figure is mislabelled, contains data that should not be plotted together, and there is no statistical testing of the hypothesis. The line marked Severe OSA is actually Moderate-Severe OSA, which means the line labelled mild and moderate OSA is probably the line for mild OSA from our data. The graph compares the OSA data collected on men and women of a rural town in Western Australia in 1990 with heavy smoking Finnish businessmen collected in 1974. Apart from obvious socioeconomic, geographical, and gender issues, the follow-up periods do not overlap (1990-2004) and (1974- ca. 1988). There are no reference categories for "No OSA" or "Never smoked" provided in their figure. Furthermore these are not independent categories because the Finns could have had OSA and some of the Australians smoked. The vertical axis has been re-drawn to maximise the apparent differences rather than giving a clear characterization of the absolute risk of death.